A driver monitoring system monitors distraction, drowsiness, and the like, of a driver while driving a vehicle to judge a factor impeding driving in advance and warn the driver of the factor impeding driving in case that the driver does an action other than the driving. This has a significant effect on preventing a traffic accident due to drowsy driving, or the like.
The driver monitoring system generally employs a technology of judging a face and a line-of-sight of the driver using a camera.
However, since an image processing technology using the camera is significantly affected by a surrounding illumination environment, a detection rate significantly changes according to day, night, weather, a position of sun, or the like.
Recently, a technology of increasing a detection rate of an image processing scheme through an illumination control has been introduced. However, since this technology acquires and uses two frames from a camera image, it is vulnerable to noise due to a time difference between the frames. Therefore, a need exists for a method for increasing a detection rate of a driver from one image.